Should a Skill Level Ever Go Down?
We built a five-step practice ladder and found an uncomfortable question at the top: what happens when someone stops practising?
Ibrahim Elkobeissi · · 5 min read ·

The ladder
Skill development in coordination work does not happen in a training room. It happens on the floor, in the moments between patients, in the correction after a difficult situation, in the quiet review at the end of a shift. A one-day course does not make a coordinator. Neither does two years without someone noticing what she does.
We use a five-step practice ladder built on the Dreyfus model of skill acquisition. The five steps are: learning, applying, consistent, fluent, and reference for colleagues. Each step describes how a coordinator relates to a practice, not how long she has worked or what she scored on a test.
- Learning. She knows the practice exists and can describe it. She has not tried it under pressure.
- Applying. She has tried it. She needs to think about it consciously while doing it.
- Consistent. She applies it without being reminded. It is part of her normal work, not a technique she is running in parallel.
- Fluent. She adapts it to situations the original description did not cover. She can explain what she changed and why.
- Reference for colleagues. She is the person others come to when they are stuck on this practice. She can demonstrate it, explain the variations, and correct someone else without making a performance of it.
Spaced confirmation, not a single test
Moving from one step to the next requires confirmation at four intervals: 7, 14, 28 and 56 days. A coordinator who applies a practice correctly on a Tuesday and again the following Tuesday, then again two weeks later, then again a month after that, has four data points. Each confirmation is a round observation, not a formal assessment. It is a note by a coordinator or a programme manager that says: I saw her do this, in this situation, in this way.
Why four? One observation proves the practice happened once. Four spaced observations suggest it has become part of how the person works. The spacing matters. A burst of four confirmations in one week shows the practice at one point in time. Four confirmations spread over eight weeks show that it held.
Promotion does not require perfection at each step. It requires that the practice is visible, that it is being used, and that the gap between confirmations is not growing.
The question at the top
Once a coordinator reaches "reference for colleagues", the ladder has nowhere to go. That is by design. The fifth step is not a rank. It is a description of function. She is the person others come to. That does not change unless she stops being that person.
And here is the uncomfortable part: what if she stops?
A coordinator at the reference step leaves for three months on leave. When she returns, the landscape has changed. New situations have emerged that she was not part of. Some of the specific techniques she mastered are now handled differently by the team. She is still experienced. She is not still the reference.
Dropping her back to step three would be wrong. It discards real knowledge and signals that absence is penalised. Keeping her at step five would also be wrong. It tells the rest of the team that the step means something it no longer means for her.
What we do instead: flag, not demote
A practice not seen for twice its confirmation interval is marked as needing a fresh look. The confirmation interval for step five is 56 days. If a coordinator has not been observed in that practice for 112 days, the system marks it. Not as a demotion. As a question: is this practice still active?
The mark opens a conversation. The coordinator and her manager decide whether the practice is still in use, whether it has changed, whether it needs a new round of confirmations. The decision is theirs. The system flags; people decide.
That principle — the system organises, people decide — is not specific to this feature. It is a rule we apply across the product. Data can tell you that a practice has not been confirmed in four months. It cannot tell you whether that matters. The person who works in that place every day can.
The open question
We have one unresolved question about the ladder, and we are keeping it open because we have not seen enough programmes run to answer it well.
Should a single level split into three separate signals?
The three signals would be: proven mastery (the four spaced confirmations, unchanged), recent practice (whether the practice has been observed within the standard interval), and evidence weight (how many total confirmations exist in the record, as a rough proxy for depth).
Separating them would let a manager see a coordinator who has deep mastery and a long record but has not been recently observed differently from a coordinator who reached the same step quickly with minimal confirmations. The first is likely rusty; the second has not yet proved the practice is durable.
The reason we have not changed the design yet is that splitting one number into three creates a new cognitive load on the manager. Three signals are not automatically more useful than one number. We are watching how programme managers use the current data before deciding what, if anything, to add.
We will decide when we have seen the question matter in practice, not before.
Reflection runs the five-step ladder for frontline coordinators: observations recorded by name, confirmations tracked at the right intervals, and the flag — not demotion — when a practice needs a fresh look. Programme managers see each coordinator's path; coordinators see only their own. See how it fits on our page for hospitals and medical centres.
Sources
- Mind Over Machine: The Power of Human Intuition and Expertise in the Era of the Computer. Free Press, 1986.
- Spaced Repetition and Learning. Cognitive Psychology Review, 2008.
Comments
Built into ROADYOM Reflection
Learn more
